Solution · Voice AI

Inbound voicebots on your PBX.

Native STT, TTS, and turn-taking on top of Asterisk. No SIP gymnastics, no per-minute, no voice-vendor lock-in.

What you replace

Voicebots without renting the voice.

Voicebot SaaS charges per minute and locks you into one vendor stack. Verbara does the opposite.

	Verbara	Vapi	Bland.ai · Retell	Pipecat (OSS)
Open-core	✓ MIT + Apache	✗ closed SaaS	✗ closed SaaS	✓ Apache
Full self-host	✓	✗	✗	✓
Asterisk-native (no SIP gymnastics)	✓ AMI/ARI	SIP gateway	SIP gateway	✗ WebRTC only
Multi-tenant	✓ Pro	✓ enterprise	✓ enterprise	✗
6 STT + 6 TTS swappables	✓ 6 STT + 6 TTS	curated	curated	✓ pluggable
OpenAI Realtime bridge	✓ direct	✗	✗	✓ direct
Smart Turn + barge-in	✓ native	✓	✓	✓

Read the code

A voicebot in 30 lines.

VoiceAgent.cs

using Verbara.Sdk.Ari;
using Verbara.Sdk.VoiceAI;
using Verbara.Sdk.VoiceAI.Stt.Deepgram;
using Verbara.Sdk.VoiceAI.Tts.ElevenLabs;

var ari = new AriClient("http://pbx:8088", "verbara", "secret");
var agent = new VoiceAgent(
  stt: new DeepgramStt(apiKey: Env("DEEPGRAM_KEY"), model: "nova-2"),
  tts: new ElevenLabsTts(apiKey: Env("ELEVENLABS_KEY"), voice: "Rachel"),
  llm: new OpenAiLlm(apiKey: Env("OPENAI_KEY"), model: "gpt-4o-mini"),
  options: VoiceAgentOptions.Default with {
    SmartTurnDetection = true,
    BargeIn = true,
  });

await ari.OnStasisStart(async (channel) => {
  await channel.AnswerAsync();
  await agent.RunAsync(channel);
});

await ari.SubscribeAsync("voicebot");

Verbara.Sdk.VoiceAI — Deepgram + ElevenLabs agent answering an Asterisk call →

FAQ · Voice AI

Builder questions.

STT: Deepgram, Google, Whisper, Azure, Cartesia, AssemblyAI, Speechmatics. TTS: ElevenLabs Flash 2.5, Deepgram Aura 2, LMNT, Azure, Cartesia, Speechmatics. Plus a direct bridge to the OpenAI Realtime API. Swappable via configuration, no recompile.
With Smart Turn detection + Deepgram Nova + ElevenLabs Flash 2.5 + barge-in: ~600ms p95 from human end-of-utterance to first TTS phoneme, measured on dedicated infra at VU 100. Exact latency depends on your infra and upstream LLM.
You need an Asterisk PBX so the SDK can receive audio (yours, a customer's, or deployed alongside Verbara). If you want voicebots without operating telephony, Tier 3+ SaaS hosted includes managed PBX.
The voicebot runs on the same runtime as the contact center, so an escalation to a human lands on a real agent using the in-browser WebRTC softphone — no desk phone, no install. And if that agent loses connectivity mid-transfer, presence liveness detects it and fires a priority callback to the customer toward the next live agent. AI voice and human voice share one queue; they are not two products taped together.

Voice AI pricing

Where do you start?

Tier 0 community is free for evaluation with the MIT SDK directly. Tier 0.5 (Pro Developer, free 30 days) for every Pro feature. Tier 1+ when you ship to single-tenant production.

See full pricing →

Voicebots without rent.

30-day signed developer license, free. MIT SDK — read it end-to-end before adopting.

Get a developer license →